The aim of this study was to empirically investigate the relationship between the age of an organization and the nature of its innovative activity in order to clarify the somewhat contradictory results of previous studies on the subject.
More specifically, three hypotheses about the effect of aging on an organization's explorative, exploitative, and overall innovative activity were formed based on previous literature on exploration, exploitation, organizational ambidexterity, aging, and innovation. These hypotheses were then statistically tested with a set of biotechnology patent data covering the modern biotechnology industry in Finland from its birth in 1973 until the year 2008.
The differences observed between explorative and exploitative patents in the data indicated that, on average, explorative patents were applied younger than the exploitative ones. This makes sense as there needs to be exploration first in order to have something to exploit. The results of series of logistic regression analyses showed that aging has a positive effect on the likelihood of an organization's innovative action to be exploitative and a negative effect of its likelihood to be explorative. The relationship of aging and overall innovative activity of an organization seemed to be so, that the likelihood of an innovative action decreases with aging which was contradictory to what was hypothesized. The analyses also showed that in addition to aging, the variable Financing (indicating if an organization had recently received private equity financing) was a significant predictor of both the type and overall likelihood of an innovative action and also the variables GDP (the yearly GDP of Finland) and Density (the overall amount of biotechnology firms) were significant in predicting the overall likelihood of an innovative action. Though, it seems that
GDP is not really a significant predictor but only appears as one due to its very strong correlation with the Density variable. Yet, although these significant predictors were identified, the regression models resulting from the analyses were not very good predictors of the phenomenon overall. Even though the coefficient of determination clearly improved when only the dedicated biotechnology firms were included in the data (ruling out the companies with biotechnology activities), they were still only 23.4 % and 8.1 % respectively for the model predicting the type of an innovative action and the model predicting the overall innovative activity of an organization. The correlation analyses conducted for all the variables used in the study supported the results of the logistic regression analyses. Only the direction of the effect the Financing variable on the dependent variables was different in the regression and correlation analyses. Some significant relationships between the dependent variable and the predicting variables in the correlation analysis also appeared as insignificant in the logistic regression analysis. The differences between the results of correlation and regression analyses, however, are not uncommon as correlation analysis considers the relationship between two variables separately but regression analysis takes into account also the effect of the other variables included in the model on the relationship.
The results of the study indicate that age affects the the type of an innovative action of an organization and the overall likelihood of the organization to conduct an innovative action so, that as an organization ages, the likelihood of the innovative action to be explorative and the overall likelihood of an innovative action decreases and the likelihood of the innovative action to be exploitative increases. Yet, the effect of aging is not very strong and it alone is a poor predictor of the phenomenon. Overall, the variables used here are not explaining the phenomenon with a very high reliability, although the 23.4 % coefficient of determination gained from the first set of logistic regression analyses can be considered sufficient for a complex phenomenon (see Ketokivi 2009: 103). In order to be able to predict the dependent variables more reliably, especially the overall innovative activity of an organization for which only a 8.1 % coefficient of determination was obtained with the best model, more variables that are able to explain the variance of the dependent variables would need to be included.
The results are in line with the the thoughts of Abernathy and Utterback (1978), Tushman and Anderson (1986) and Sørensen and Stuart (2000) whose ideas were also the basis of the hypothesis formation in this study. Abernathy and Utterback (1978) claimed that an organization's innovative behavior changes from radical (explorative) to incremental (exploitative) over time.
Tushman and Anderson (1986) concluded that new entrants are most often responsible of competence-destroying (explorative) innovation as the incumbent firms are the ones most often responsible of competence-enhancing innovations (exploitative) in an industry. Sørensen and Stuart (2000) discussed organizational competence and environmental fit that together cause older organizations to stay with their existing areas of expertise (exploitation) and young organizations to more often search for new areas (exploration) due to the increase in organizational competence and decrease in environmental fit in a
53 changing environment with time. All these ideas are in line with explorative innovative actions decreasing and exploitative innovative actions increasing with age as was shown here. However, unlike the ideas of Sørensen and Stuart (2000) and Tushman and Anderson (1986) suggest about the innovative activity increasing with age as was hypothesized here also, the results indicated that the likelihood of an innovative action actually decreases with age. Sørensen and Stuart (2000) suggested that the growing organizational competence leads to growing rate of innovations as an organization ages and Tushman and Anderson (1986) claimed that the competence-destroying innovations that are most often initiated by new entrant firms are rarer than the competence-enhancing innovations that are most often initiated by incumbent firms. The reason for this surprising result could be the nature of the biotechnology industry that causes a bias in the data. The industry is largely characterized by specialized firms small in size (Allansdottir et al. 2002) and due to the difficulties of reaching economic profitability and lack of resources, many small biotechnology companies have been short-lived as they have been sold to bigger companies at a young age (Ruutu 1990). As the data here consists solely of biotechnology organizations, this results to most of the organizations in the data being rather young. As there are more young than old organizations in the data, the result of young organizations being more likely to innovate than older ones could partly arise faultily from the fact that the relative portion of young organizations in the data is greater. Had the analysis been conducted with a data consisting of the CWBA (companies with biotechnology activities) firms only, the result would most probably have been different as the CWBA firms contain more old firms and the difference between the young and the old firms in the data would not have been as great. This was not done, however, as the purpose of the study was to focus on the biotechnology industry and as the CWBA firms tend to operate in several industries, their actions are not as bound to the biotechnology industry as they are for the firms that operate solely in this field. The closer investigation of this relationship of the age and overall innovative activity of an organization is an interesting topic for further research.
The previous empirical studies on the relationship of aging and the nature (explorative or exploitative) of an organization's actions are contradictory. The results of the previous studies are mostly in line with the results of the analyses conducted here. As was found here, Sørensen and Stuart (2000) found that young organizations are more likely to explore than their older counterparts and older organizations, on the other hand, are more exploitative of nature.
Also Coad and Guenther (2013) as well as Choi and Phan (2014) found the explorative actions to decrease with age. Voss and Voss (2013) found this negative relationship in product domain only. The result of the likelihood of overall innovative activity declining with age is supported by the results of Coad and Guenther (2013) who found a similar relationship and the results of Withers et al. (2011) who found this negative relationship only when the level of innovation capabilities of the organization were low. However, there are also studies in which a relationship of the opposite direction has been found. Coad and Guenther (2013) found also the exploitative actions to decrease with age and Voss and Voss (2013) found that this is the situation in market domain.
Regarding the overall innovative activity, Sørensen and Stuart (2000) and Shi and Zhu (2014) found an increase in the innovative activity as an organization ages. Also Withers et al. (2011) found a similar relationship when the level of innovation capabilities of the organization were high. Few of the studies found also other features of the relationship of firm age and the nature of its behavior that are not in line with the results here. Dunlap–Hinkler et al. (2010) found that here is no correlation between firm age and exploration, Xie and O'Neill (2014) found an U-shaped relationship between firm age and exploitation, and Huergo and Jaumandreu (2004) found no relationship between the overall innovative activity and firm age on a general level, but they also mentioned that this depends on the industry in question. In addition, Chen (2014) and Coad and Guenther (2013) found the relationship of firm age and innovative activity to be wave shaped.
The main reason for the results differing from some of the results of previous studies can, again, be found in the nature of the biotechnology industry. As Huergo and Jaumandreu (2004) also note, the changes in innovative activity of an organization over time are industry dependent. As most of the previous studies have been conducted on different industries, some of them in settings very far from the biotechnology field in such industries as the non-profit professional theaters (Voss & Voss 2013), it is not surprising that the results are somewhat different from the ones in this study. If only the three previous studies that are conducted in the settings most similar to the one in here are considered, it is evident that the results actually are not so different.
These three studies are Sørensen and Stuart's study from 2000 conducted on biotechnology (and semiconductor) industry, the study of Dunlap-Hinkler et al.
from 2010 conducted on pharmaceutical industry, and the study of Shi and Zhu from 2014 conducted on pharmaceutical (and IT) industry. As mentioned, the results of Sørensen and Stuart (2000) on the nature of an organization's innovative behavior are similar to the ones here. However, regarding the overall innovative activity, the results differ from the ones of Sørensen and Stuart (2000) and Shi and Zhu (2014), but this is most likely due to the data of this study containing mostly young organizations as was discussed before.
Another dissimilarity in the results is that, unlike here, Dunlap-Hinkler et al.
(2010) found no correlation between firm age and explorative activity. They also claim that the previous results of Sørensen and Stuart (2000) about age affecting the innovative activity of an organization are caused by firm size (that usually correlates with firm age) and not age. This is most likely not the case, however, since the vast majority of the DBFs are small firms (Allansdottir et al.
2002: 37) and according to the results here, age seems to have a significant effect on the nature of innovative activity even when the data is limited to DBFs only.
The differences are likely to occur due to other reasons. They could be due to the difference in the sample population as Dunlap-Hinkler et al. (2010) did not view the whole biotechnology industry but focused on pharmaceutical companies only. As the effect of age was significant, but not very strong here, it would make sense that it could turn insignificant in certain populations.
As mentioned, the ability of the logistic regression models to predict the value of the dependent variables increased when the data was limited to only
55 DBFs (compared to the data containing both DBFs and CWBAs). This is, first of all, due to the division of the patents as the function of firm age. As is shown by Figure 1, the frequencies of patents applied at each age are rather steady in the younger end, but very scattered in the older end with several ages at which no patents have been applied. When the CWBAs are excluded from the data, this scattered end is left out as the DBFs only have patents applied at ages 0–22 (see Table 4), which makes it easier to fit a model in the data. Another reason could be the increased homogeneity of the data as the CWBAs operating in other industries too are not as bound to the biotechnology industry as the DBFs that are focusing only on biotechnology.
The coefficients of determination that were 23.4 % and 8.1 % for the models predicting the nature of innovative activity and overall innovative activity. This means that the variables included in the models explain rather small portion of the variation of the dependent variables. As Age also was not the most significant predictor in the model, the conclusions made from the effect of age should be kept moderate. An organization's age does have a statistically significant effect on its innovative behavior, but this effect is not very strong.
7.2 Contribution, limitations and future directions
This study contributes to the firm level studies of organizational ambidexterity as well as to the previous studies of the effects of aging and an organization's innovative behavior. The study provides evidence of the effects of aging on innovative behavior in the context of the modern biotechnology industry in Finland.
The fact that the size of the organizations was not controlled in the study could be seen as a major limitation of the study as the effect of aging has been criticized to actually be caused by firm size (see Dunlap-Hinkler et al. 2010). As said, however, the DBF firms are generally small in size. According to Allansdottir et al. (2002: 37), 90 % of the European DBFs are small or micro sized firms. Even though the size was not directly controlled, the fact that age was a significant predictor of innovative behavior in a population containing only DBFs that are generally of similar size, it can be said that the effect of age is independent of the effect of size. The size was not directly controlled here due to the simple reason that the size information was not available for many of the organizations in the data.
Another point of possible criticism is the use of patent data as the measure of innovative behavior. As mentioned, patent data has been criticized due to the patenting behavior and the value of patents differing between industries (Levin et al. 1987; Gambardella et al. 2008). However, in the biotechnology industry, intellectual property is comprehensively protected with patents (Levin et al.
1987: 786) which make them a valid measure of the technological activities in this industry (Belderbos et al. 2010: 874).
A true limitation of this study is the somewhat artificial division of the patents to explorative and exploitative ones. As each organization was considered separately, the possibility of knowledge and experience moving
from one firm to another after, for example, a merger or an acquisition is ruled out. In reality, in many cases some or all the know-how of a predecessor firm is likely to move to the successor firm causing some of the actions categorized here as explorative ones to actually be exploitative of nature. Due to the practical limitations, this could not be considered, however. In most cases, there was no information available on to what extent (if at all) the knowledge or know-how of a firm was transferred to its successor (for example for an acquiring firm). As said, this leads to the division of the patents to be partially artificial. Yet, as the possible know-how inherited from a predecessor firm could not have been reliably measured here, the division of the patents to explorative and exploitative ones is as reliable as was possible to achieve.
The results of the study provide important information on the relationship of firm age and innovative behavior. The results, however, are not generalizable to other industries as they are and further research on the phenomenon is needed. The generalizability of the results is limited due to the effect of aging varying between industries as is indicated by the empirical studies conducted on the subject (see for example Huergo and Jaumandreu 2004). As Sørensen and Stuart (2000) discuss, it is not only the firm-level aging process, but also the evolutionary process on the industry level, that affects innovative behavior. In order to generalize the results, they need to be considered together with other empirical results from different industries and contexts.
In addition to the differences in the effect of age on innovative behavior between industries, another topic for future research to deal with is the question of what exactly are the causes that reliably explain an organization's innovative behavior. This topic has already gained interest and there are empirical studies on the topic (see for example Sørensen & Stuart, 2000;
Dunlap-Hinkler et al., 2010), but due to the complexity and context dependence of the phenomenon, further studies are needed in order to understand it and to make universal conclusions. As this research shows, firm age, received private equity financing, and the number of competitors in the industry are important predictors of the phenomenon in the biotechnology industry. Even though it is already shown that sufficient funding is important for succeeding in the biotechnology industry (Schienstock & Tulkki, 2001) and that equity financing has a positive effect on innovation rates (Kortum & Lerner, 2000), the role of financing in relation to the nature of the innovative behavior would be an interesting future avenue for research. Out of all the variables used in this study, the received private equity financing turned out to be the most important predictor of the nature of an organizations innovative behavior. Financing, however, was a dummy variable here, including only the information whether or not an organization had recently received private equity financing (due to the lack of information on the exact amount in many cases). To more thoroughly examine the effect of financing on an organization's innovative behavior, the role of both private and public financing should be further investigated.
57
REFERENCES
Abernathy, W. J. & Utterback, J. M. 1978. Patterns of Industrial Innovation.
Technology Review June/July, 41–47.
Academy of Finland. 2002. Biotechnology in Finland: Impact of Public Research Funding and Strategies for the Future. Publications of the Academy of Finland 11/02. Helsinki: Academy of Finland.
Allansdottir, A., Bonaccorsi, A., Gambardella, A., Mariani, M., Orsenigo, L., Pammolli, F. & Riccaboni, M. 2002. Innovation and Competitiveness in European Biotechnology. Luxembourg: Office for Official Publications of the European Communities.
Barnett, W. P. 1990. The Organizational Ecology of a Technological System.
Administrative Science Quarterly 35 (1), 31–60.
Barron, D. N., West, E. & Hannan, M. T. 1994. A Time to Grow and a Time to Die: Growth and Mortality of Credit Unions in New York City. American Journal of Sociology 100 (2), 381–421.
Belderbos, R., Faems, D., Leten, B. & Van Looy, B. 2010. Technological Activities and Their Impact on the Financial Performance of the Firm: Exploitation and Exploration within and between Firms. Journal of Product Innovation Management 27, 869–882.
Brüderl, J. & Schüssler, R. 1990. Organizational Mortality: The Liabilities of Newness and Adolescence. Administrative Science Quarterly 35 (3), 530–
547.
Burgelman, R. A. 1991. Intraorganizational Ecology of Strategy Making and
Burgelman, R. A. 1991. Intraorganizational Ecology of Strategy Making and